Deformable mirrors are often used in the field of adaptive optics. For example, phase distortions in a signal may be sensed by a wavefront sensor and these distortions may be corrected for by an adaptive mirror. Such adaptive mirrors may be employed in numerous fields, including:                imaging, for example adaptive mirrors are used in astronomy to improve the resolution of earth-based telescopes that are otherwise affected by atmospheric distortions;        laser sensing, where the amount of laser light that can be delivered onto a target is significantly increased by using an adaptive mirror to correct for atmospheric distortions—this enables either better information to be obtained or objects to be identified at a greater range; and        laser generation, where an adaptive mirror can be used intracavity within a high power laser to counter the thermal blooming that can be otherwise induced by the high concentration of laser light inside the cavity.        
Bimorph deformable mirrors have been proposed as low cost adaptive mirrors. The two main operational parameters of a bimorph mirror are its bandwidth and its stroke. Bandwidth determines how quickly the mirror can be deformed and hence, for example, how quickly the mirror can respond to the variations in atmospheric turbulence. Stroke corresponds to the maximum displacement of the mirror when deformed and this determines, for example, the level of turbulence that can be corrected. Ideally, both bandwidth and stroke would be maximised. However, conventional designs mean that there is a reciprocal relationship between these two parameters, and one parameter can only be improved at the expense of the other. Therefore, to date, designers have always looked for ways to improve either the resonant frequency or the stroke independently from each other.
Conventionally, deformable mirrors are supported rigidly around their edge, for example an annular ring overlapping the periphery of the mirror is used to hold the mirror firmly in position. Such an arrangement benefits from being simple yet rugged. However, it has an inherent disadvantage in that it creates a dead space around the mirror's edge. This corresponds both to the area of the mirror held firmly under the annular ring and also to the adjacent area. This is because the useable area of the deformable mirror (the active area) must bend to adopt a desired profile, for example either a concave or a convex shape. The annular area between the active area and the annular ring must bend in the opposite sense and so forms an area of inflexion that has undesirable optical properties. Hence the active area occupies only a central portion of the whole mirror. This is illustrated in FIGS. 1 to 3, albeit with the deformation of the mirror exaggerated for the sake of clarity.
One way of alleviating this problem is to clamp the mirror at only three positions, such that the mirror edge can twist. However, this arrangement is to the detriment of ruggedness and the twisting introduces unwanted distortions leading to an optically inferior performance.
It is also known that for smaller mirrors, bonded piezoelectric elements (e.g. bimorph deformable mirrors) are preferred due to their relatively low cost. Such mirrors provide an adequate balance between bandwidth and stroke. However, the balance between bandwidth and stroke is especially important when looking to make larger mirrors e.g. mirrors with active apertures greater than 10 to 15 cms. In order to keep the resonant frequency and thus the bandwidth of the mirror constant, the thickness of the substrate must also increase. For the larger mirrors this will adversely affect the minimum curvature available from the mirror. For this reason, larger mirrors have historically been zonal mirrors. The fact that the substrate is supported by a large number of actuators means that the resonant frequency, and therefore bandwidth, is no longer directly linked to the mirror diameter. However, the overriding issue with this type of deformable mirror is the cost. Although there are a number of different actuator technologies available, none of them are cheap. This makes large mirrors expensive because as many as 300 actuators may be required. For a bonded piezoelectric element mirror, although a large piezoelectric element will be more expensive than a smaller one, the cost differential will not be as great. A second issue is that it is not always possible to place discrete actuators as close to each other as required because of their fairly large size.
Against this background, the Applicant has filed a UK patent application (No. 0309976.9, 30 Apr. 2003 filing date) directed to a deformable mirror comprising a reflective surface provided on a substrate and a layer of deformable material attached to the substrate that is operable to deform the mirror and wherein the substrate is supported by an actuator that is operable to deform the mirror. The actuator can be used to provide the basic deformation required of the mirror (e.g. focus), while the deformable material can be used to provide fine tuning of the mirror shape.
Additionally, against this background, the Applicant has filed a UK patent application (No. 0309979.3, 30 Apr. 2003 filing date) directed to a deformable mirror holder that simply supports the mirror uniformly around the edge.